distiller summary
TRANSCRIPT
---------------------------------_.....,-1306 S.A. MEDICAL JOURNAL 22 June 1974
Asymptomatic Bacteriuria in Diabetes MellitusB. I. JOFFE, H. C. SEFTEL, L. A. DISTILLER
SUMMARY
A study was undertaken to determine the frequency ofasymptomatic bacteriuria among 100 ambulant diabeticpatients attending a diabetic outpatient clinic. At the sametime, we assessed the reliability of the Uricult dip-slidemethod for detecting urinary bacterial growth. Significantbacteriuria occurred in 9% of the total diabetic group,largely attributable to the high prevalence in elderlydiabetic women. Important characteristics of the affecteddiabetic patients included infection, commonly withEscherichia coli, frequently associated pyuria, gooddiabetic control with normal renal function and a highrecurrence rate of bacteriuria after treatment. The dipslide method compared quite favourably with the laboratory culture method, but it failed to detect 2 out of 8cases with significant bacteriuria.
S. Afr. Med. J., 48, 1306 (1974).
Whether patients with diabetes mellitus have an increasedsusceptibility to urinary tract infections remains a controversial issue. Much of the conflicting literature hasbeen reviewed by Thomsen.' A recent investigation indicated that true bacteriuria in the diabetic occurred almostexclusively in females, and especially in the older agegroups.' In the present study we re-evaluate the problem ina representative group of diabetic outpatients, and assessthe usefulness of the Uricult dip-slide (Orion Laboratories,Helsinki) as a reliable screening technique. The dip-slidemethod has been satisfactorily employed in a variety ofother clinical situations, including paediatric practice;protein calorie malnutrition; urological units' and obstetricpatients!
PATIENTS AND METHODS
Urine samples were obtained from 100 ambulant patientsattending the Diabetic Clinic at the Johannesburg Hospital. None of the patients admitted to symptoms of renalinfection' or urinary incontinence at the time, and nonehad recently been catheterised; they were otherwise unselected. The group included 60 females, 40 of whom wereover 60 years of age (Table I).
Since elderly women in general may be prone to bacteriuria,' a control group of 36 ambulant non-diabeticfemale patients over the age of 60 years, who attended
Carbohydrate-Lipid Research Unit, Department of Medicine,University of the Witwatersrand, Johannesburg
B. I. JOFFE, M.B. B.CH., M.R.C.P.H. C. SEFTEL, B.SC., M.B. B.CH., D.MED.L. A. DISTILLER, M.B. B.CH., F.C.P. (S.A.)
Date received: 25 February 1974.
the Outpatient Department for a variety of chronic m«dical disorders, were also studied. They, too, had no symptoms of recent kidney infection or urinary incontinence.
A clean voided midstream urine sample was collectedfrom each subject, afteJ;" perineal cleansing, in. a sterileuniversal container. Uricult dip-slides (one side of whichis coated with MacConkey's medium and' the other withCLED agar) were dipped into the freshly voided specimens and then stood, in their sterile containers,. at roomtemperature for 24 hours. (We chose this method ·sincereports'" claim equal reliability with formal incubation at3rC, and also for convenience.) The urine samples' weresent without delay to the bacteriology laboratory for concurrent examination. This consisted of rnicroscopy andinoculation of culture plates used as a routine, similar incomposition to the media coating the dip-slides. In 30patients (14 diabetics and 16 controls) only the laboratoryculture result was available for analysis, due to a telllporaryinterruption of dip-slide supplies.
Degrees of bacterial growth on the dip-slides and plateswere assessed as showing (a) no growth; (b) no significantgrowth (10" - 10' coloniesjml) and (c) significant growth(> 10' coloniesjml). A urinary leucocyte count of morethan 4000jrnl (centrifuged specimen) indicated significantpyuria. In patients yielding positive urine cultures, information about the chemical composition of the urine,blood sugar control and renal function at the time ofstudy was obtained; appropriate antibiotic therapy wasgiven for 2 weeks and a follow-up urine sample examinedafter 3 months.
RESULTS
Incidence of Bacteriuria in Diabetic Patients(Table I)
Significant bacterial growth occurred in 9% of patients,according to the culture plate results. The elderly femalediabetic group was mainly responsible, providing 8 of the9 'positive' cases and showing a group prevalence of 20%.Compared with the 2,8 % incidence of bacteriuria in nondiabetic elderly females, a statistically significant difference(x' 5,38; P<0,05) emerged.
Correlation between the Two Methods of UrineCulture (Fig. 1)
The correlation between the laboratory plate and thedip-slide techniques for determining urinary bacterialgrowth in the 106 subjects (86 diabetics and 20 nondiabetic elderly females) is summarised in Fig. 1. Theshaded blocks indicate agreement of clinical significance
>
n Junie 1974 S.-A. MEDIESE TVDSKRIF 1307
TABLE I. INCIDENCE OF BACTERIURIA IN DIABETIC PATIENTS AND NON-DIABETIC ELDERLY FEMALE CONTROLS
Incidence of bacteriuria
Group
TotalMales ...Females below 60 years of ageFemales above 60 years of age'Control' non-diabetic females above 60 years of age
Mean age (range)
57 (16-82)54 (16-77)42 (24-59)68 (61-82)72 (61-88)
No.
10040204036
No. %9 91 2,50 08 20*1 2,8
• Significantly higher <P<O,05) than the non-diabetic female control group <as well as the other two diabetic subgroups).
LABORATORY PL~TE METHOD
No growth No sig. growth Sig. growth
w~
oZ:J:UW....WC...J
Cl'Ia.CFig. 1. Correlation between laboratory plate method andUricult dip-slide technique for detecting urinary bacterialgrowth. The numbers in the blocks represent cases and theshaded blocks indicate agreement of clinical significance(see text).
(i.e. no discrepancy or a discrepancy not involving>10' bacterial growth in one test and a lesser growth inthe other). Identical results were obtained in 85 of 106
cases (80%); clinically ummportant differences· (no growthin one test and a W' - 10" growth in the other) in 18 instances (17%); and a serious discrepancy arose in 3 patients(3%). The dip-slide technique failed to detect 2 of the 8positive cases identified in the laboratory where pairedcomparisons were made; in one instance the laboratorygrew Staphylococcus and in the other an Escherichia coli.(Two positive diabetic urines were examined in the laboratoryonly.)
Characteristics of the 9 Diabetic Patients with Significant Bacteriuria (Table ID
Several salient points emerge from the tabulated data.E. coli was the most frequent infecting organism; pJ1lriaalmost always accompanied the bacteriuria; both glycosuria and proteinuria were usually absent at the time ofurine sampling; diabetic control appeared to be 'good'(random blood glucose below 150 mg/lOO ml) in themajority of patients; renal function (assessed by bloodurea level) was normal in all except one instance; andsystemic hypertension occurred in one-third of the cases.Five of the 9 patients still had significant bacteriuria whenre-examined after a 3-month interval.
DISCUSSION
This study confirms the impression that true bacteriuria inthe diabetic is mainly confined to elderly diabetic women.Furthermore, the prevalence of bacteriuria among them
TABLE 11. CLINICAL AND LABORATORY FEATURES IN 9 DIABETICS WITH SIGNIFICANT BACTERIURIA
Diabetic Blood urea
Case Sex Age Organism Pyuria Glycosuria Proteinuria control* Hypertension (mg/l00 ml) 3-mo. follow-up
1 F 74 E. coli Yes Nil Nil Good Present 40 E. coli still'present
2 F 73 Proteus Yes Nil Nil Good Present 34 Clear urine
3 F 63 E. coli Yes Nil Nil Moderate Absent 88 Recurrence with Proteus
4 F 61 E. coli Yes Mild Nil Good Absent 22 E. coli still present
5 F 68 E. coli Yes Nil Nil Good Absent 36 Clear urine
6 F 61 E. coli Yes Nil Mild Good Present 38 Clear urine
7 F 65 E. coli Yes Nil Nil Good Absent 32 E. coli still present
8 F 62 Klebsiella No Nil Nil Moderate Absent 27 Klebsiella still present
9 M 46 Staph. Yes Nil Nil Good Absent 31 Clear urine
pyogenes
• Good = random blood sugar <150 mg/loo ml; moderate = random blood sugar between 150 m~ and 250 m~/100 ml.
1308 S.A. MEDICAL JOURNAL 22 June 1974
Q
was significantly greater than that of elderly non-diabeticfemales similarly investigated.
Criticisms can be levelled at our findings, e.g. that thenumber of patients in the various subgroups was relativelysmall and that the method of urine collection was notideal. Repeated sampling or suprapubic aspiration ofurine might have been more desirable.' The answer tothis is that the study was a practical one to screen a representative cross-section of diabetic clinic attenders withoutundue inconvenience to patients or staff, and employingcurrently used outpatient procedures.
Reasons for the enhanced susceptibility of elderly diabetic females to asymptomatic bacteriuria are unclear.Factors such as inadequate diabetic control or impairedrenal function do not appear to be critical, nor does theadvanced age of the patients per se. Retrospective inquiryfailed to reveal evidence of analgesic abuse as a cause.'·Underlying structural abnormalities of the urinary tractmight be important, and Batalla et al.' fOUD_d a high incidence of neurogenic bladders with intravenous pyelography. This may also partly explain the frequency of recurrent bacteriuria on follow-up examination. Unfortunately information concerning pyelography was not availablein our patients. Renal micro-angiopathy has been implicated as an important precursor of pyelonephritis in thediabetic,n but has not played an important role in thisstudy, as significant proteinuria was absent in all 9 bacteriuric cases.
Finally, concerning the usefulness of the Uricult dipslide technique as a screening procedure for significantbacteriuria in a diabetic population, our findings providesome support for what has been found in other clinical
situations. A preliminary dip-slide test will reduce thelaboratory workload by screening off negative urines; inthose cases yielding significant growth, further investigations by standard methods can then be instituted. Onedisturbing aspect, however, was the failure of the methodto detect 2 positive cases confirmed in the laboratory.The probable explanation is that incubation at room temperature retarded bacterial growth-especially in the patientwith Gram-positive bacteriuria.". Incubation of the slidesat 3rC may reduce the likelihood of false negative results.'"
We should like to thank Boehringer Mannheim for theUricult dip-slides.
This study was supported by grants from the South AfricanMedical Research Council and the Witwatersrand UniversityCouncil.
REFERENCES
I. Thomsen, A. C. (1965): The Kidney in Diabetes Melliws, pp. 106 - 128.Copenhagen: Munksgaard.
2. BataIJa, M. A., Balodimos, M. C. and Bradley, R. F. (1971): Diabetologia, 7, 297.
3. Arneil, G. c., McAllister, T. A. and Kay, P. (1970): Lancet, 1, 119.4. Buchanan, N. (1971): S. Afr. Med. J., 45, 320.
. 5. Jackman, F. R., DarrelJ, J. H. and Shackman, R. (1973): Brit. Med.J., 1, 207.
6. Bailey, M. J., Neary, J. T. F. and Notelovitz, M. (1972): S. Afr. Med.J., 46, 1323.
7. De Wardener, H. E. (1963): The Kidney, 2nd ed., p. 258. London:Churchill.
8. McMillan, S. A. (1972): Lancet, 2, 452.9. Stamey, T. A. (1972): Urinary In/ections, p. 4. Baltimore: Williams &
Wilkin.s.10. Kingsley, D. P. E., Goldberg, B., Abrahams, C., Meyers, A. I.,
Furman, K. I. and Cohen, I. (1972): Brit. Med. I., 4, 656.11. Bruns, W. (1972): Acta diabet. lat., 9, 46.12. McAllister, T. A., Arneil, G. C., Barr, W. and Kay, P. (1973):
Nephron, 11, 111.13. Wille, L, Schorer, K. and Bickel, H. (1970): Lancet, 2, 105.